The child and household fixed effects specifications in Table 8 provide evidence that child
fostering can lead to a Pareto improvement in school enrollment for all children involved: the host
siblings in the receiving family, the biological siblings remaining behind in the sending family, and
the foster child. This Pareto improvement is largest for the youngest children. The dataset used
in Table 8 contains observations from 640 host siblings, 994 biological siblings, 316 foster children
and 470 children from non-fostering households measured over 3 years, with the reference group in
the regressions being the children from the non-fostering households.
In column 1, I control for the child’s type in a child fixed effects regression and find that the
host siblings, biological siblings, and foster children are, respectively, 1.9, 1.2, and 1.1 percent more
likely to be enrolled after the fostering compared to children from non-fostering households. The
regression controls for age effects by including dummy variables indicating if the child is young or
middle aged, but in this regression in which children of all ages are pooled together, the coefficients
are not statistically significant. The coefficient estimates for host siblings, biological siblings, and
foster children in column 1 are roughly equal which is consistent with the child fixed effects results
in Table 6, where foster children are no better off following the fostering relative to either host or
biological siblings.
The earlier tables showed age was significantly correlated with enrollment outcomes for these
children, and in columns 2, 3, and 4, I estimate child fixed effects regressions for subsets of children.
For households that exchange a young child, the results indicate that relative to non-fostering
household children, all young children are much better off after the fostering. Host siblings increase
enrollment by 23.9 percent, biological siblings by 11.3 percent, and foster children by 21.5 percent
relative to the non-fostering household children, and all coefficients are significant at the 1 percent
level. For older children aged 8 to 11, the positive impact of fostering relative to children from
non-fostering households diminishes, with biological siblings’ enrollment increasing by 4.1 percent
19